ingenius super-elevation idea?

[DeadheadGreg]

So I'm just looking at a bunch of D&RGW pics from this one guys AMAZING site, and I see a picture of a train heading into a curve in the open country.  The engines are just onto the curved part, with the rest of the train behind it, on the straight part, and you can see the engines banking.

This got me thinking:  How the **** does the prototype make super-elevated curves?  You can't just pile up more ballast and expect the ties to stay angled, can you?  And then that would be nearly impossible to keep consistent, right?

SO, all of a sudden this idea just hit me:  why not just use bigger rail for the outside rail, or smaller rail for the inside?  This way, you can just lay track regularly, without having to go through the trouble of laying a line of ties all the way down the outer edge of a curve before you begin laying any track.

And changing rail height wouldn't matter...  because if you are on the outside of the curve, you could just use smaller rail on the inside rail, and keep the outside rail the same, so there wouldn't be any visual difference.  Or if you're on the inside of a curve, just use bigger rail on the outside, because thats the furthest away from you.

Since super-elevation isn't like, a big huge obvious thing, like if the tracks were at a 45 degree angle, I think that simply changing the rail height on one rail would be enough to trick your eyes and mind into thinking that both rails are the same size, and that the track is indeed banked.

 

genius?  stupid?  already done?  i'm curious to see what you guys think

[cacole]

Real railroads don't do it that way, so why try to do it in the model world?  This would work only if you're using smaller rail for all of the rest of the layout.  I know of no supplier of HO scale rail taller than code 100, so in order to elevate to that code you would have to start with code 70, then step to code 83, then to 100.  How are you going to make a smooth transition from one code to the next?  Filing the rail top smooth would be difficult.  It's much simpler to put something under the outer edge of the ties and use the same code rail all the way around.

[markpierce]

This idea has no appeal.  It would be difficult to transition the heights of the rail.   (You want me to file a 2-foot section to a consistent slope?  That's more work than any other method I'm aware of even if I was capable).  Appearance wise, it would still bother me.  The amount of transition will probably be inappropriate (transition from code 70 to 83 in HO scale would create an 11" superelevation which is something in the order of 2 times the amount wanted).

Mark

[grizlump9]

 "You can't just pile up more ballast and expect the ties to stay angled, can you?  And then that would be nearly impossible to keep consistent, right?"

that is exactly how they do it.  also the ballast shoulder on the outside rail is more substantial to keep the track from sliding out due to centrifical force.   curves always require more attention from the mow dept. to keep the rail elevation constant and the ballast tamped tight.  

also, rails become curve worn on the gauge side of the outside rail and frequentlly have to be replaced.

this is why railroads try to use the widest curves possible.  not just to maintain train speed but also to minimize track problems.

on our model railroads we can get away with shimming the outside of the flex-track and covering up with ballast.  superelevation can come back to bite you though.  long freight trains may tend to stringline on such trackwork.  if you have enough of a bank to enable passenger trains to maintain speed you have enough to cause trouble with heavy freights pulling hard on the curve.  it is always a trade off.

the DRGW had this problem a lot as well as dealing with the rail "walking" or slipping downgrade and bunching up at the bottom of the hill.

grizlump

[tomikawaTT]

In observing the prototype I have seen several different methods of achieving superelevation:

• Bank the entire roadbed.  The ballast depth is constant, but the unballasted 'shoulder' is higher on the outside of the curve.
• Bank the ties by varying the depth of ballast on a level roadbed.
• Raise the outside rail by placing tapered wooden wedges under the outside rail at every tie.  The ties themselves are level.  (Seen on the Pelham Bay Line elevated structure of the New York Subway System)

 

In fact, the only way I have never seen is the use of taller steel on the outer rail.  If this was ever attempted, the outside rail would have to be 12-14 inches high.  PRR's infamous 155#/yd rail is only 8 inches high.  Not to mention that the rollover stress on such tall rail would be a metallurgist's nightmare.

Chuck (Modeling Central Japan in September, 1964)

[markpierce]

My early 20th-century railway engineering textbook says that sometimes railroads raised the outer rail and lowered the inner rail.  This eleminated extra work for the locomotive since it didn't have to literally lift one side of the train for the several-inch elevation if superelevation was created by only raising the outer rail.  The drop of the lowered inner rail compensated for any rise on the outer rail.

Mark

[doctorwayne]

If you use open grid or L-girder benchwork, superelevation is easy to add, including the vertical easements into and out of the curve.

I use 3/4" plywood as a sub-roadbed, but any similar-type material, or even spline roadbed should also work.  Install the straight roadbed on either side of the curve by fastening the risers to the joists - leave the last riser beyond both ends of the curve unfastened.  Install risers to the underside of the curved roadbed, but don't fasten them to the benchwork just yet.  If your curve is on a grade, as most of mine are, raise the roadbed through the curve to the proper height, then mark a pencil line on each riser which corresponds to the top of the benchwork to which it will eventually be fastened.  If the track through the curve is to be level, adjust the risers accordingly, then make the lines.  Next, choose the riser closest to the mid-point of the curve, raise it to the proper height, then push the bottom end of the riser towards the outside of the curve.  Re-align the height line on the riser so that its inner end corresponds to the top of its benchwork member (the height line will be tilted, with the end on the outside of the curve somewhat above the benchwork).  I've found that the best way to establish the amount of superelevation is by placing a train on the curve, then adjusting the off-set of the bottom of the riser until it "looks right".  I use a C-clamp while I'm making the visual adjustments, then, when I'm satisfied with the appearance, that mid-point riser is screwed to the benchwork.  Because the roadbed is torsionally flexible, each riser on either side of the mid-point will now be off-set from the vertical, to diminishing degrees, as the distance from the mid-point increases.  Working from the mid-point of the curve, carefully raise each riser so that the inside end of the height mark aligns with the top of the benchwork to which it will be fastened, making sure to not change the angle of off-set, then screw the risers to the benchwork.  I did all of mine with the trackwork in place.

It's difficult to see in this photo, but the track at the far end of the bridge is on a broad S-curve, with super-elevation throughout - it's very enjoyable watching trains roll past as they heel into one curve, then to the opposite side on the next.   The near end of the bridge is also super-elevated.

 

Wayne   

[markpierce]

Wayne, that's a great looking bridge......................

One method railroads used to put superelevation on a non-ballasted bridge was to make the bridge ties wedge-shaped.

[steemtrayn]

doctorwayne

It's difficult to see in this photo, but the track at the far end of the bridge is on a broad S-curve, with super-elevation throughout - it's very enjoyable watching trains roll past as they heel into one curve, then to the opposite side on the next.     

Can you make us a video?

[doctorwayne]

steemtrayn

Can you make us a video?

 

Sorry, but not with my camera.

Wayne

[doctorwayne]

doctorwayne

steemtrayn

Can you make us a video?

 

Sorry, but not with my camera.

Wayne

 

Your request intrigued me enough to attempt a video, and, after several attempts, here are the best two (and don't think that my use of "best" implies that they're good, as they're not). 

This first one was shot with the camera on a tripod, which was sitting atop another part of the layout.  I think that the effects of the superelevation are noticeable, but you'll have to be the judge of that.

This second one took several attempts, as the flatcar on which the camera was riding either kept derailing due to the weight of the camera being concentrated at one end of the car, or, as the car entered the reverse curve, the shift in superelevation caused the flatcar to tip as the camera slid to the opposite side.  Finally, I used a piece of double-sided tape to secure the camera to the deck, resulting in this: (Watch out for that last sign post!)

Despite the poor image quality, I think that this one shows the superelevation better.

Wayne   

[tstage]

Sounds like a lot of extra work to me.  You would need to transition between the changes in rail height for each rail.

I find strips of .01", .02", and .03" styrene directly underneath the outside rail to work quite well and it's very simple to apply.  Remember the Albert Einstein adage:

"Make everything as simple as possible, but not simpler."

Tom

[reklein]

Looks to me like the video attemt was worthwhile. Good enuf quality and perfect length to illustrate the effect. BILL

[doctorwayne]

Thanks, Bill. 

Tom, the beauty of my method is that the vertical easement of the individual rails is introduced automatically.  The curves shown in the videos range from about a 33" radius on the initial one (which is on the end of a peninsula, turning the track back slightly more than 180*)  to around 48" through the "S" bend.  Since I chose to not drop the inside rail, its height remains constant (within the 2.8% grade, of course), while the outer rail rises and falls gradually on either side of the mid-point of the curve, and continues to do so even when the "outside" rail changes sides as it progresses through the "S" bend.  No mathematical formulae, no shims of varying thickness, no fuss, no muss.    Of course, this method works only where the sub-roadbed is a separate construction from the rest of the layout - table-top layouts, either on foam or directly on plywood can't benefit from this method.

Wayne

[Scarpia]

Thank you for the video, Dr. Wayne - as always, your advice helps.  Sorry about that sign post though!

[doctorwayne]

Scarpia

Thank you for the video, Dr. Wayne - as always, your advice helps.  Sorry about that sign post though!

Yeah, I removed the ones from the near-side of the bridge, but forgot about those at the other end.  That run was the first that made it to the end of the track, as the camera kept sliding off the flatcar as the train passed through the curves.  I finally used some double-sided tape to keep it in place.  The superelevation, as you can see, isn't that severe, but with the weight of the camera concentrated at the rear of the flatcar, but hanging over both edges, it didn't take much to cause the car to tilt slightly atop its trucks. 

Where the curve transitions from a left-hand one to a right-hand one, it's almost 5' to the concrete floor, with the track only 1" from the edge of the layout.   I often run heavy coal trains here, with "live" loads, but, so far, no mishaps.  The combination of heavy trains, live loads, and the big drop encourage operator attention at all times, with slow and careful running.

Wayne

[ndbprr]

I had superelevated curves on one railroad and just used extra ties under the outer rail and ties to raise the outside rail.  Having just installed eight interior doors and casings if I did it today I would use wood shims available at any big box store like Loews or Home Depot.  They are two or three dollars for about twenty five of them and all have the same angle of slope. It would be very easy to use them and gradate the elevation uniformly.  Once all are in place cutting them off with a utility knfe is very easy to do as they are soft pine.

[CSX Robert]

cacole

Real railroads don't do it that way, so why try to do it in the model world?

While I don't think this idea would work very well(if you can live with the visual difference of the rail height, the hardest part would be the transition from one to the other), I think that this is one of the least valid reasons not to do it. These are, after all, models and we do often have to find ways of doing things that are not prototypical. Generally speaking, real railroads do not glue their track and ballast down, don't have a single electric motor in the middle of the locomotive turning worm gears and spur gears in each truck and picking up current from the rails, have plastic ties, foam or cork roadbed, etc., etc(notice that I did say "generally," because for each one of these, someone could probably find a prototype instance of it).

markpierce

The amount of transition will probably be inappropriate (transition from code 70 to 83 in HO scale would create an 11" superelevation which is something in the order of 2 times the amount wanted)

Actually, that the difference between code 70 and code 83 would be a scale 1.1" [ (83 - 70) / 1000 * 87 = 1.131].

[markpierce]

CSX Robert

cacole

Real railroads don't do it that way, so why try to do it in the model world?

While I don't think this idea would work very well(if you can live with the visual difference of the rail height, the hardest part would be the transition from one to the other), I think that this is one of the least valid reasons not to do it. These are, after all, models and we do often have to find ways of doing things that are not prototypical. Generally speaking, real railroads do not glue their track and ballast down, don't have a single electric motor in the middle of the locomotive turning worm gears and spur gears in each truck and picking up current from the rails, have plastic ties, foam or cork roadbed, etc., etc(notice that I did say "generally," because for each one of these, someone could probably find a prototype instance of it).

markpierce

The amount of transition will probably be inappropriate (transition from code 70 to 83 in HO scale would create an 11" superelevation which is something in the order of 2 times the amount wanted)

Actually, that the difference between code 70 and code 83 would be a scale 1.1" [ (83 - 70) / 1000 * 87 = 1.131].

Of course you're correct (dang that decimal).  Now, 1.1" hardly seems worthwhile. I'd think about two to four times that much elevation would be needed to bother at all where trains run at speed and more in line with prototypical superelevation according to my railway engineering textbook.

Mark